Skip to content


Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time


cracken crate cracken version cracken documentation cracken total downloads

Cracken is a fast password wordlist generator, Smartlist creation and password hybrid-mask analysis tool written in pure safe Rust (more on talk/). Inspired by great tools like maskprocessor, hashcat, Crunch and πŸ€— HuggingFace's tokenizers.

What? Why? Woot??

At DeepSec2021 we presented a new method for analysing passwords as Hybrid-Masks exploiting common substrings in passwords by utilizing NLP tokenizers (more info on talk/).

Our method splits a password into its subwords instead of just a characters mask. HelloWorld123! splitted into ['Hello', 'World', '123!'] as these three subwords are very common in other passwords.

Hybrid Masks & Smartlists

  • πŸ“„ Smartlists - Compact & representative subword lists created from passwords by utilizing NLP tokenizers
  • 🎭 Hybrid-Mask - A representation of a password as a combination of wordlists & characters (e.g. ?w1?w2?l?d)

Analyzing RockYou Passwords with Smartlists & Hybrid-Masks:

Top25 Hybrid Masks from RockYou

full table here

Cracken πŸ™ is used for:

  • βœ… Generating Hybrid-Masks very VERY FAST πŸ¦Έβš‘πŸ’¨ (see performance section)
  • βœ… Building Smartlists - compact & representative list of subwords from given passwords files (using πŸ€— HuggingFace's tokenizers)
  • βœ… Analyzing passwords for their Hybrid-Masks - building statistics for better password candidates (again very fast)

Possible workflows with Cracken:


  1. Generate wordlist candidates from a hybrid mask - e.g. cracken -w rockyou.txt -w 100-most-common.txt '?w1?w2?d?d?d?d?s'
  2. You can pipe the passwords Cracken generates into hashcat, john or your favorite password cracker


  1. Create a Smartlist from existing passwords - cracken create
  2. Analyze a passwords list of plaintext passwords - cracken entropy
  3. use most frequent Hybrid-Masks to generate password candidates fast - cracken generate -i hybrid-masks.txt

For more details see Usage section

Getting Started

download (linux only currently): latest release πŸ”—

for more installation options see installation section

run Cracken:

generate all words of length 8 starting with uppercase followed by 6 lowercase chars and then a digit:

$ cracken -o pwdz.lst '?u?l?l?l?l?l?l?d'

generate words from two wordlists with year suffix (1000-2999) <firstname><lastname><year>

$ cracken --wordlist firstnames.txt --wordlist lastnames.lst --charset '12' '?w1?w2?1?d?d?d'

create a Smartlist of size 50k from subwords extracted from rockyou.txt

$ cracken create -f rockyou.txt -m 50000 --smartlist smart.lst

estimate the entropy of hybrid mask of the password HelloWorld123! using a smartlist

$ cracken entropy -f smart.lst 'HelloWorld123!'

hybrid-min-split: ["hello", "world1", "2", "3", "!"]
hybrid-mask: ?w1?w1?d?d?s
hybrid-min-entropy: 42.73
charset-mask: ?l?l?l?l?l?l?l?l?l?l?d?d?d?s
charset-mask-entropy: 61.97


As of writing this, Cracken is probably the world's fastest wordlist generator:

bechmarks results

Cracken has around 25% increased performance over hashcat's fast maskprocessor thats written in C.

Cracken can generate around 2 GB/s per core.

more details on benchmarks/ πŸ”—

Why speed is important? A typical GPU can test billions passwords per second depending on the password hash function. When the wordlist generator produces fewer words per second than the cracking tool can handle - the cracking speed will degrade.

Hybrid-Masks Analysis Performance

Cracken uses A* algorithm to analyze passwords very fast. it can find the minimal Hybrid-Mask of passwords file at rate of ~100k Passwords/sec (cracken entropy -f words1.txt -f words2.txt ... -p pwds.txt)


install Cracken or compile from source

Download Binary (Linux Only Currently)

download latest release from releases πŸ”—

Build From Source (All Platforms)

Cracken is written in Rust and needs rustc to get compiled. Cracken should support all Platforms that Rust support.

installation instructions for cargo πŸ”—

there are two options building from source - installing with cargo from (preferred) or compiling manually from source.

1. install from (preferred)

install with cargo:

$ cargo install cracken

2. build from source

clone Cracken:

$ git clone

build Cracken:

$ cd cracken
$ cargo build --release

run it:

$ ./target/release/cracken --help

Usage Info

$ cracken --help
Cracken v1.0.0 - a fast password wordlist generator 

    cracken [SUBCOMMAND]

    -h, --help       Prints help information
    -V, --version    Prints version information

    generate    (default) - Generates newline separated words according to given mask and wordlist files
    create      Create a new smartlist from input file(s)
                Computes the estimated entropy of password or password file.
                The entropy of a password is the log2(len(keyspace)) of the password.
                There are two types of keyspace size estimations:
                  * mask - keyspace of each char (digit=10, lowercase=26...).
                  * hybrid - finding minimal split into subwords and charsets.

For specific subcommand help run: cracken <subcommand> --help

Example Usage:

  ## Generate Subcommand Examples:

  # all digits from 00000000 to 99999999
  cracken ?d?d?d?d?d?d?d?d

  # all digits from 0 to 99999999
  cracken -m 1 ?d?d?d?d?d?d?d?d

  # words with pwd prefix - pwd0000 to pwd9999
  cracken pwd?d?d?d?d

  # all passwords of length 8 starting with upper then 6 lowers then digit
  cracken ?u?l?l?l?l?l?l?d

  # same as above, write output to pwds.txt instead of stdout
  cracken -o pwds.txt ?u?l?l?l?l?l?l?d

  # custom charset - all hex values
  cracken -c 0123456789abcdef '?1?1?1?1'

  # 4 custom charsets - the order determines the id of the charset
  cracken -c 01 -c ab -c de -c ef '?1?2?3?4'

  # 4 lowercase chars with years 2000-2019 suffix
  cracken -c 01 '?l?l?l?l20?1?d'

  # starts with firstname from wordlist followed by 4 digits
  cracken -w firstnames.txt '?w1?d?d?d?d'

  # starts with firstname from wordlist with lastname from wordlist ending with symbol
  cracken -w firstnames.txt -w lastnames.txt -c '!@#$' '?w1?w2?1'

  # repeating wordlists multiple times and combining charsets
  cracken -w verbs.txt -w nouns.txt '?w1?w2?w1?w2?w2?d?d?d'

  ## Create Smartlists Subcommand Examples:

  # create smartlist from single file into smart.txt
  cracken create -f rockyou.txt --smartlist smart.txt

  # create smartlist from multiple files with multiple tokenization algorithms
  cracken create -t bpe -t unigram -t wordpiece -f rockyou.txt -f passwords.txt -f wikipedia.txt --smartlist smart.txt

  # create smartlist with minimum subword length of 3 and max numbers-only subwords of size 6
  cracken create -f rockyou.txt --min-word-len 3 --numbers-max-size 6 --smartlist smart.txt

  ## Entropy Subcommand Examples:

  # estimating entropy of a password
  cracken entropy --smartlist vocab.txt 'helloworld123!'

  # estimating entropy of a passwords file with a charset mask entropy (default is hybrid)
  cracken entropy --smartlist vocab.txt -t charset -p passwords.txt

  # estimating the entropy of a passwords file
  cracken entropy --smartlist vocab.txt -p passwords.txt

cracken-v1.0.0 linux-x86_64 compiler: rustc 1.56.1 (59eed8a2a 2021-11-01)
more info at:

Generate Subcommand Usage Info

$ cracken generate --help
(default) - Generates newline separated words according to given mask and wordlist files

    cracken generate [FLAGS] [OPTIONS] <mask> --masks-file <masks-file>

    -h, --help       
            Prints help information

    -s, --stats      
            prints the number of words this command will generate and exits

    -V, --version    
            Prints version information

    -c, --custom-charset <custom-charset>...    
            custom charset (string of chars). up to 9 custom charsets - ?1 to ?9. use ?1 on the mask for the first charset

    -i, --masks-file <masks-file>               
            a file containing masks to generate

    -x, --maxlen <max-length>                   
            maximum length of the mask to start from

    -m, --minlen <min-length>                   
            minimum length of the mask to start from

    -o, --output-file <output-file>             
            output file to write the wordlist to, defaults to stdout

    -w, --wordlist <wordlist>...                
            filename containing newline (0xA) separated words. note: currently all wordlists loaded to memory

            the wordlist mask to generate.
            available masks are:
                builtin charsets:
                ?d - digits: "0123456789"
                ?l - lowercase: "abcdefghijklmnopqrstuvwxyz"
                ?u - uppercase: "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
                ?s - symbols: " !\"\#$%&'()*+,-./:;<=>?@[\\]^_`{|}~"
                ?a - all characters: ?d + ?l + ?u + ?s
                ?b - all binary values: (0-255)
                custom charsets ?1 to ?9:
                ?1 - first custom charset specified by --charset 'mychars'
                wordlists ?w1 to ?w9:
                ?w1 - first wordlist specified by --wordlist 'my-wordlist.txt'

Create Smartlist Subcommand Usage Info

$ cracken create --help  
Create a new smartlist from input file(s)

    cracken create [FLAGS] [OPTIONS] --file <file>... --smartlist <smartlist>

    -h, --help       Prints help information
    -q, --quiet      disables printing progress bar
    -V, --version    Prints version information

    -f, --file <file>...                         input filename, can be specified multiple times for multiple files
        --min-frequency <min_frequency>          minimum frequency of a word, relevant only for BPE tokenizer
    -l, --min-word-len <min_word_len>            filters words shorter than the specified length
        --numbers-max-size <numbers_max_size>    filters numbers (all digits) longer than the specified size
    -o, --smartlist <smartlist>                  output smartlist filename
    -t, --tokenizer <tokenizer>...               tokenizer to use, can be specified multiple times.
                                                 one of: bpe,unigram,wordpiece [default: bpe]  [possible values: bpe, unigram, wordpiece]
    -m, --vocab-max-size <vocab_max_size>        max vocabulary size

Entropy Subcommand Usage Info

$ cracken entropy --help

Computes the estimated entropy of password or password file.
The entropy of a password is the log2(len(keyspace)) of the password.

There are two types of keyspace size estimations:
  * mask - keyspace of each char (digit=10, lowercase=26...).
  * hybrid - finding minimal split into subwords and charsets.

    cracken entropy [FLAGS] [OPTIONS] <password> --smartlist <smartlist>...

    -h, --help       Prints help information
    -s, --summary    output summary of entropy for password
    -V, --version    Prints version information

    -t, --mask-type <mask_type>              type of mask to output, one of: charsets(charsets only), hybrid(charsets+wordlists) [possible values: hybrid, charset]
    -p, --passwords-file <passwords-file>    newline separated password file to estimate entropy for
    -f, --smartlist <smartlist>...           smartlist input file to estimate entropy with, a newline separated text file

    <password>    password to




Cracken is under active development, if you wish to help below is this the partial roadmap for this project. Feel free to submit PRs and open issues.